SM

SmolLM2 360M Instruct Heretic

SmolLMGGUF by mradermacherUpdated 2026-02-21
Download on Hugging Face
Parameters362M
Context8K
Layers32
Architecturellama
Smallest quant0.2 GB
Quick answer

SmolLM2 360M Instruct Heretic needs about 1.1 GB of memory at Q4_K_M (a 0.3 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 12 quants, from 0.2 GB (Q2_K) to 0.7 GB (F16).

SmolLM2 360M Instruct Heretic – a decensored 360 M parameter Llama model

SmolLM2 360M Instruct Heretic is a 0.362 B parameter Llama‑based model with a 8192 token context window. It is a decensored version of SmolLM2‑360M‑Instruct, released under the apache‑2.0 license. The model contains no safety filters and can produce offensive content.

Good at

  • Fast inference on low‑memory GPUs
  • Compact size for edge deployment
  • Supports 8192 token context

Limits

  • No safety filters – use responsibly
  • No vision or audio capabilities
  • Requires 8 k context for full performance

Best for

  • Local chatbot prototypes
  • Rapid prototyping of instruction‑following models
  • Testing decensoring techniques

Who it’s for. Developers and researchers who need a lightweight, instruction‑following Llama model without safety filters. Suitable for experimentation and edge deployment where memory is limited.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K0.2 GB 1.1 GB
Fits Get file →
Q3_K_S0.2 GB 1.1 GB
Fits Get file →
IQ4_XS0.2 GB 1.1 GB
Fits Get file →
Q3_K_M0.2 GB 1.1 GB
Fits Get file →
Q3_K_L0.2 GB 1.1 GB
Fits Get file →
Q4_K_S0.3 GB 1.1 GB
Fits Get file →
Q4_K_M0.3 GB 1.1 GB
Fits Get file →
Q5_K_S0.3 GB 1.1 GB
Fits Get file →
Q5_K_M0.3 GB 1.1 GB
Fits Get file →
Q6_K0.4 GB 1.2 GB
Fits Get file →
Q8_00.4 GB 1.2 GB
Fits Get file →
F160.7 GB 1.6 GB
Fits Get file →

For 8 GB VRAM, Q4_K_M (0.27 GB) or Q4_K_S (0.26 GB) are good choices. If you have 16 GB VRAM, Q6_K (0.37 GB) or Q8_0 (0.39 GB) provide higher quality. The F16 quant (0.73 GB) gives the best precision but needs more memory.

What GPU do you need for SmolLM2 360M Instruct Heretic?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: the best quant that fits is F16 (0.7 GB file, about 1.6 GB in use).

Which SmolLM2 360M Instruct Heretic quant should you choose?

Q4_K_M is the usual starting point for SmolLM2 360M Instruct Heretic: it balances file size and quality well. Here it is a 0.3 GB file.

The Q8_0 file is 0.4 GB, about 1.4 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.

If memory is tight, the smallest file is Q2_K at 0.2 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturellama
Parameters362M
Layers32
Trained context8,192 tokens
KV cache (fp16)40 KB per token

Source

Quantized bymradermacher
FamilySmolLM
Licenseapache-2.0
Downloads (HF)1,045
Last updated2026-02-21

Frequently asked questions about SmolLM2 360M Instruct Heretic

How much VRAM does SmolLM2 360M Instruct Heretic need?

About 1.1 GB at Q4_K_M with an 8K context window. That is the 0.3 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run SmolLM2 360M Instruct Heretic on an 8 GB GPU?

Yes, with F16: the 0.7 GB file needs about 1.6 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of SmolLM2 360M Instruct Heretic GGUF?

It depends on the quant: Q2_K is 0.2 GB, Q3_K_S is 0.2 GB, IQ4_XS is 0.2 GB, Q3_K_M is 0.2 GB, Q3_K_L is 0.2 GB, Q4_K_S is 0.3 GB and 6 more.

What context length does SmolLM2 360M Instruct Heretic support?

The GGUF header lists a trained context length of 8,192 tokens (8K). The KV cache grows with the context you use: about 0.3 GB at 8K and 0.3 GB at 8K.

Where can I download SmolLM2 360M Instruct Heretic GGUF?

The files are on Hugging Face in the mradermacher/SmolLM2-360M-Instruct-Heretic-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does SmolLM2 360M Instruct Heretic use?

Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run SmolLM2 360M Instruct Heretic on 8 GB VRAM?

Yes, use Q4_K_M or Q4_K_S quantizations; each requires about 1.1 GB memory at 8k context.

Does this model support vision or audio input?

No, it is a text‑only Llama model with no vision or audio capabilities.

What license applies to this model?

The model is released under the apache‑2.0 license.

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More SmolLM models

All SmolLM →

Other tiny models

All tiny models →